Modeling multivariate distributions for bayesian optimal inference

[Cat]

Sorry, I was just posting it

Pymc3 gives shifted result compared to analytical bayes solution

Hello I’m (very) new to pymc3, it’s a great tool that I’d like to learn more about. I’m trying to understand the syntax and methods by testing pymc3 inference on simple distributions where I can compute the bayesian inference result analytically (for example a single Beta function) The pymc3 result is a bit different than the bayes optimal. Does anyone know why this might happen? # with pm.Model() as model: # Define the Beta beta1_ = pm.Beta('beta1', alpha=5.1, beta…